NVIDIA/cuda-rust

Issues somebody may already be working on are listed last. List everything by date

• [BUG]: cuTile: PointerTile method chain fails with "return type is missing a compiled tile type"

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 48/100

  NVIDIA/cuda-rust#1596 · Opened Oct 21, 2026

  Maintainers usually reply within 2 days

• [BUG]: f32::to_int_unchecked fails with "invalid input program" in a kernel

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 1-2 days Newbie friendliness 55/100

  NVIDIA/cuda-rust#1595 · Opened Oct 21, 2026

  Maintainers usually reply within 2 days

• [BUG]: float min/max lower to setp+selp instead of min.f32/max.f32

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 53/100

  NVIDIA/cuda-rust#1594 · Opened Oct 21, 2026

  Maintainers usually reply within 2 days

• chore: post cutile-rs cleanup

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 42/100

  NVIDIA/cuda-rust#1573 · Opened Oct 20, 2026

  Maintainers usually reply within 2 days

• [BUG]: simt Empty and Value are Send without checking what they hold

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  bug

  Difficulty 2/5 1-3 hours Newbie friendliness 70/100

  NVIDIA/cuda-rust#1568 · 1 comment · Opened Oct 20, 2026

  Maintainers usually reply within 2 days

• cuda-core: nvls_all_reduce example references are missing a path

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 1/5 Under an hour Newbie friendliness 88/100

  NVIDIA/cuda-rust#1519 · Opened Oct 15, 2026

  Maintainers usually reply within 2 days

• cutile: improve the autotune example

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  enhancement TBD

  Difficulty 3/5 1-2 days Newbie friendliness 45/100

  NVIDIA/cuda-rust#1517 · Opened Oct 14, 2026

  Maintainers usually reply within 2 days

• feat(cuda-device): add PTX cache-control streaming and eviction qualifiers for global memory

  May be free again A pull request for this issue was closed without being merged. Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cuda-oxide device-apis enhancement intrinsics

  Difficulty 4/5 3-5 days Newbie friendliness 42/100

  NVIDIA/cuda-rust#1514 · Opened Oct 13, 2026

  Maintainers usually reply within 2 days

• API design: organize shared CUDA host types under semantic namespaces

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 5/5 Over a week Newbie friendliness 8/100

  NVIDIA/cuda-rust#1505 · Opened Oct 13, 2026

  Maintainers usually reply within 2 days

• cuda-core: distinguish CUDA devices and contexts in the shared host API

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 5/5 Over a week Newbie friendliness 8/100

  NVIDIA/cuda-rust#1504 · Opened Oct 13, 2026

  Maintainers usually reply within 2 days

• cuda-core: define boolean storage and element padding/packing policy

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cuda-oxide cutile-rs interop runtime

  Difficulty 5/5 Over a week Newbie friendliness 8/100

  NVIDIA/cuda-rust#1503 · 1 comment · Opened Oct 13, 2026

  Maintainers usually reply within 2 days

• feat(cargo-oxide): select extra-rustflags by target triple

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 48/100

  NVIDIA/cuda-rust#1481 · Opened Oct 11, 2026

  Maintainers usually reply within 2 days

• [BUG]: tileiras 13.4 emits misaligned f16 loads with reshape/broadcast on sm_89

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 42/100

  NVIDIA/cuda-rust#1479 · 1 comment · Opened Oct 10, 2026

  Maintainers usually reply within 2 days

• #[unroll] over the clone budget leaves an array-indexing loop rolled and moves the array to .local, with no way to raise the budget

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 5/5 Over a week Newbie friendliness 45/100

  NVIDIA/cuda-rust#1478 · Opened Oct 10, 2026

  Maintainers usually reply within 2 days

• Adding a kernel renames other kernels' shared-memory statics, so their PTX changes

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 43/100

  NVIDIA/cuda-rust#1477 · 1 comment · Opened Oct 10, 2026

  Maintainers usually reply within 2 days

• Ship prebuilt kernel artifacts from a library crate to users on stable Rust

  May be free again @0130w claimed this 9 days ago, and no pull request is open. Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  build-related cargo-oxide cuda-oxide documentation enhancement host-apis

  Difficulty 4/5 3-5 days Newbie friendliness 18/100

  NVIDIA/cuda-rust#1464 · 3 comments · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• [BUG]: *const [T] kernel param gets 1 host arg but the kernel expects 2

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  bug

  Difficulty 3/5 1-2 days Newbie friendliness 72/100

  NVIDIA/cuda-rust#1455 · 1 comment · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• fix: ignore load_ptr_tko padding when no mask is given

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs post-migration-pr

  Difficulty 3/5 1-2 days Newbie friendliness 44/100

  NVIDIA/cuda-rust#1440 · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• feat(cuda-core): borrow external streams as CudaStream

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs post-migration-pr

  Difficulty 3/5 1-2 days Newbie friendliness 20/100

  NVIDIA/cuda-rust#1439 · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• Add bytecode reader. Allows cutile-rs to run existing cuda-tile kernels

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs post-migration-pr

  Difficulty 4/5 3-5 days Newbie friendliness 25/100

  NVIDIA/cuda-rust#1436 · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• fix: FP8 tests are expected to fail on ampere+ada

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs post-migration-pr

  Difficulty 1/5 Under an hour Newbie friendliness 15/100

  NVIDIA/cuda-rust#1435 · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• Cutile frontend

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs post-migration-pr

  Difficulty 4/5 3-5 days Newbie friendliness 20/100

  NVIDIA/cuda-rust#1434 · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• Extract cutile-macro concepts into cutile-expand and cutile-syn-utils crates

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs post-migration-pr

  Difficulty 4/5 3-5 days Newbie friendliness 25/100

  NVIDIA/cuda-rust#1433 · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• feat(cutile): add reduce_xor, reduce_and, and reduce_or bitwise reductions

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs post-migration-pr

  Difficulty 1/5 Under an hour Newbie friendliness 20/100

  NVIDIA/cuda-rust#1432 · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• feat(cuda-core): add zero-copy PinnedHostMapping and DeviceAllocation bridge

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs post-migration-pr

  Difficulty 4/5 3-5 days Newbie friendliness 15/100

  NVIDIA/cuda-rust#1431 · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• feat: expose CUDA device attributes

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs post-migration-pr

  Difficulty 4/5 3-5 days Newbie friendliness 25/100

  NVIDIA/cuda-rust#1430 · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• feat: add CUDA graph capture modes

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs post-migration-pr

  Difficulty 4/5 3-5 days Newbie friendliness 25/100

  NVIDIA/cuda-rust#1429 · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• feat: model full debug and add TileIR JIT overrides

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs post-migration-pr

  Difficulty 4/5 3-5 days Newbie friendliness 25/100

  NVIDIA/cuda-rust#1428 · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• feat(cutile): legalize mixed FP8xFP4 mmaf_scaled for Tile IR 13.3

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs post-migration-pr

  Difficulty 4/5 3-5 days Newbie friendliness 25/100

  NVIDIA/cuda-rust#1427 · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• feat(cutile): add raw-offset memory ops with generic forwarding support

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs post-migration-pr

  Difficulty 5/5 Over a week Newbie friendliness 20/100

  NVIDIA/cuda-rust#1426 · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• feat(cutile): support scalar specialization overrides

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs post-migration-pr

  Difficulty 4/5 1-2 days Newbie friendliness 25/100

  NVIDIA/cuda-rust#1425 · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• fix(compiler): resolve the element type of scalar pointer tiles

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs post-migration-pr

  Difficulty 2/5 1-3 hours Newbie friendliness 1/100

  NVIDIA/cuda-rust#1424 · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• feat: host conversions for the narrow float storage types

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs post-migration-pr

  Difficulty 4/5 3-5 days Newbie friendliness 25/100

  NVIDIA/cuda-rust#1423 · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• feat: decode Tile bytecode into modules

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs post-migration-pr

  Difficulty 5/5 Over a week Newbie friendliness 20/100

  NVIDIA/cuda-rust#1422 · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• Fix host copy source lifetime

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs post-migration-pr

  Difficulty 4/5 1-2 days Newbie friendliness 35/100

  NVIDIA/cuda-rust#1421 · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• Add MemPool reuse policy configuration and trim_to

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs post-migration-pr

  Difficulty 4/5 3-5 days Newbie friendliness 25/100

  NVIDIA/cuda-rust#1420 · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• doc: add infra/aws-gpu Terraform bootstrap and AGENTS.md runbook

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs post-migration-pr

  Difficulty 4/5 3-5 days Newbie friendliness 20/100

  NVIDIA/cuda-rust#1419 · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• Cleanup zip tuples

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs post-migration-pr

  Difficulty 3/5 1-2 days Newbie friendliness 35/100

  NVIDIA/cuda-rust#1418 · Opened Oct 8, 2026

  Maintainers usually reply within 2 days

• Reactor scanner: one fixed spin count cannot serve both long waits and fast completions

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs

  Difficulty 5/5 Over a week Newbie friendliness 35/100

  NVIDIA/cuda-rust#1413 · 1 comment · Opened Oct 7, 2026

  Maintainers usually reply within 2 days

• [BUG]: cutile 0.3.1 JIT fails with 'unable to resolve generic type \u0026 f32' for kernels using \u0026Tensor / \u0026mut Tensor parameters

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  bug cutile-rs

  Difficulty 4/5 3-5 days Newbie friendliness 42/100

  NVIDIA/cuda-rust#1412 · Opened Oct 7, 2026

  Maintainers usually reply within 2 days

• CompileIQ integration: search and apply Advanced Control Files on top of cutile::tune

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs

  Difficulty 5/5 Over a week Newbie friendliness 25/100

  NVIDIA/cuda-rust#1411 · Opened Oct 7, 2026

  Maintainers usually reply within 2 days

• Check placement ignores control dependence: hoisted checks constrain executions that never reach the access

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs

  Difficulty 5/5 Over a week Newbie friendliness 35/100

  NVIDIA/cuda-rust#1410 · 1 comment · Opened Oct 7, 2026

  Maintainers usually reply within 2 days

• Validate optimization hint values against Tile IR constraints

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs

  Difficulty 4/5 3-5 days Newbie friendliness 52/100

  NVIDIA/cuda-rust#1409 · 1 comment · Opened Oct 7, 2026

  Maintainers usually reply within 2 days

• Track cross-process JIT compilation stampede on cold cache misses

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs

  Difficulty 5/5 Over a week Newbie friendliness 35/100

  NVIDIA/cuda-rust#1408 · 1 comment · Opened Oct 7, 2026

  Maintainers usually reply within 2 days

• Add execute_warmup: runtime launch warmup

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs

  Difficulty 5/5 Over a week Newbie friendliness 30/100

  NVIDIA/cuda-rust#1406 · Opened Oct 7, 2026

  Maintainers usually reply within 2 days

• Surface remaining CUDA 13.3 Tile IR API

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs enhancement

  Difficulty 5/5 Over a week Newbie friendliness 25/100

  NVIDIA/cuda-rust#1405 · Opened Oct 7, 2026

  Maintainers usually reply within 2 days

• Grout prefill is much slower than paper artifact on RTX 5070 Ti, and referenced commit 4631fb01629fc3bc36ec4302b0bf3636f23e1453 is not found

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs perf

  Difficulty 4/5 3-5 days Newbie friendliness 48/100

  NVIDIA/cuda-rust#1404 · Opened Oct 7, 2026

  Maintainers usually reply within 2 days

• Make with_global_device_context state explicit (Uninitialized | Available | Borrowed)

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs

  Difficulty 5/5 Over a week Newbie friendliness 35/100

  NVIDIA/cuda-rust#1403 · 1 comment · Opened Oct 7, 2026

  Maintainers usually reply within 2 days

• Integration with CubeCL

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs

  Difficulty 5/5 Over a week Newbie friendliness 30/100

  NVIDIA/cuda-rust#1402 · 1 reaction · Opened Oct 7, 2026

  Maintainers usually reply within 2 days

• Improve dtype/cast coverage

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  cutile-rs

  Difficulty 4/5 3-5 days Newbie friendliness 42/100

  NVIDIA/cuda-rust#1401 · Opened Oct 7, 2026

  Maintainers usually reply within 2 days